RM E53129

RESEARCH MEMORANDUM

EFFECTS OF ADDITIVES ON PRESSURE LIMITS OF FLAME

PROPAGATION OF PROPANE-AIR MIXTURES

By Frank E. Belles and Dorothy M. Simon

Lewis Flight Propulsion Laboratory Cleveland, Ohio

NATIONAL ADVISORY COMMITTEE FOR AERONAUTICS

WASHINGTON December 18,1953

https://ntrs.nasa.gov/search.jsp?R=19930088034 2018-05-06T10:14:13+00:00Z

NATIONAL ADVISORY CO- FOR AXEiOMUTICS

RESEARCH MEMORANDUM

PROPAGATION OF PROPAIW-AIR MlxTuREs '

By Frauk E. B e l l e s and Dorothy M. Simon

Seven addltives i n 0.5-volume-percent concentration w e r e studied for their effects on the low-pressure lfmits of flame propagation of propane-air mixtures. The lMts were measured i n a flame tube of new design. Mixtures containing approxlmately 2 t o 8 percent propane by volume were studled. The limft curves were without lobes on the rich side and were closely related to quenching-distance data measured by the flash-back of a Bunsen flame.

The data were analyzed by means of the experimental curves and the Le Chatelier law governing the flammability limits of mixed fuels.

Ethyl nitrate and chloropicrin were found t o be definite promoters of flame propagation i n r i c h mixtures. Chloropicrin and methyl bromide inhibited propagation in lean mixtures; it was concluded that the effect i s chemical and that these Etaaitives do not ac t merely as i ne r t gases. None of the additives promoted. flame propagation in l ean mix- tures more than could be explained by the contrfbution of the additive t o the to ta l fue l in the mixture. Methyl bromide increased the minimmu pressure f o r flame propagation and wa6 the only additive that had an appreciable effect on the minimum. Carbon disulfide had a large in- hibitory effect on flame propagation in lean mixtures, as defined by deviations f r o m the requirements of Le Chatelier's l a w .

It has long been considered important t o study catalysts for the combustion of fuels. Both positive and negative catalysts are of practical Fmportance: the positive type because they may increase the heat-release rate or widen the range of stable burning, the negative type because they may act as fire-extinguishing agents.

2 NACA RM E53129 . There are experimental data that indicate the existence of such

ca ta ly t ic e f fec ts . In the slow ( i . e . , flameless) combustion of hydro- cerbons, the addition of formaldehyde or acetaldehyde to the mixture is found t o reduce or to eliminate the induction period of the reaction ( re f . 1). I n these cases, the additive acts in a positive manner; that is , it promotes the oxidation. On the other hand, hydrogen added t o mixtures af propane and higher hydrocarbons i n oxygen may ac t as a negative catalyst; under some circmtances, the hydrogen completely inhibits the slow combustion of these fuels (ref. 1). Many other ex- amples m i g h t be c i ted. CD M

0 m In the case of f a s t burning, striking effects of additives are not

so numerous. The only agents that have been put to wide pract ical use are te t raethyl lead t o inhibi t engine knock, various halogen-containing compounds to extinguish f ires, and dopes t o improve the performance of Diesel fuels. Many materials have been tested fo r promoting effects on burning velocity; however, only slight changes were observed, and these were cons-fstent wlth the calculated changes i n equilibrium active- particle concentrations in the flames (ref. 2). The influences of various agents on the composition limtts of flammability a t atmospheric pressure have also been examined. In one such study, the cases i n which the lean limit was broadened could be explained by the added heat- release due t o the burning of the admtive (ref. 3). Some definite promoting effects on the rich limit were found i n the same work. Definite inhibiting effects on the composition limits are shorn by some halogen-containing compounds. Methyl iodide, for example, narrows the limits more than would be expected if the additive were merely an in- e r t gas ( refs . 3 and 4).

The experbents reported in references 2 t o 4 were carried out a t atmospheric pressure. There is no certainty that the resul ts would be the same at reduced pressures. In view of the importance of low- pressure burning, particularly in jet-engine combustors, a study of the effects of several selected additives on the low-pressure limits of flame propagation i n propane-air mixtures was undertaken a t the NACA Lewis laboratory. Low-pressure burning was chosen for study because recent work a t the Lewis laboratory has improved the understanding of pressure limits (refs. 5 and 6 ) ; it was anticipated that it would be possible to evaluate the resul ts so a8 t o distinguish between the var- ious possible types of additive action.

The effects of seven additives on the pressure limits of propane- air mixtures are reported herein. Each additive m s chosen because it had been reported t o have some effect on other combustion properties or because of general interest. All the t es t s were made i n a tube 3.73 centimeters i n diameter; the effects of tube diameter .on pressure limits are described in reference 5. The approximate range of propane concen- trations studied was 2 t o 8 percent by volume. In all the experiments,

NACA RM E53129 3

the additive constituted 0.5 percent by volume of the t o t a l mixture of propane, air, and additive. Propane was used because of the ease of handling and i t s genera3 slmKhrity in combustton properties t o other hydrocarbons. The resul ts of the investigation are interpreted with the aid of the Le Chatelier or mixture rule governing the flammability l imits of mixtures (ref. 7 ) .

Experimental

Apparatus. - The apparatus was basically the same as that used t o study the effect of tube diameter on the pressure lWts of propane-air mixtures (ref. 5). As in reference 5, a capacitance spark w a s used as the ignition source. Three modifications were made i n order to car ry out the present experhents: (1) A tank was provided for the preparation

liquid additives could be distilled, w a s attached t o t h e vacuum l ine; (3) a new design of t&e was incorporated.

% P and storage of addltive-air mixtures; (2) a g h s s appendix, from which

3 5

rd

The flame tube used in this invest igat ion resembled those described in reference 5 in t ha t It consisted of an Ignition section 8.7 centi- meters in &lameter and 20 centlrneters long attached t o a nazrower propa- gation section. The previously described tubes w e r e of all-glass con- struction, and the two sections were smoothly joined. It is reported in reference 5 that, i n a propane-air mixture of a given composition, the flame initiated in the ignit ion section either propagated into and throughout the length of the narrow tube or w a s extinguished a t i t s en- trance. In thi's way, it was found that the quenching distance was e q d to t he tube diameter at the pressure limit f o r propagation of flame in- t o t he tube. It is believed that more precise llmits would be obtained with an abrupt transition from the ignition section t o the propagation section, in place of the mare gradual taper that was present in the one- piece tubes.

The flame tube w~bs accordingly assenibled i n three parts (f ig. 1). The upper end of the ignition section -8 waxed fnto an ann- groove i n the brass adapter. A tapered hole w-as centered Fn the adapter t o r e - ceive the matching taper ground on the lawer end of the propagation sec- tion. !The taper joint was necessaryto prevent the tvibe from sliding into the ignition section when the apparatus was evacuated.

The propagation section itself was a precision-bore, heavy-wall glass tube 3.73 centimeters i n diameter and 50 centimeters low. Inas- much as the flames in these experiments either propagated the lengkh of the t&e o r were extinguished at i ts mouth, it was considered unnecessary t o use a t ~ e 100 centhe ters long.

4 NACA RM E53129

Preparation of propane-air-additive mixtures. - The propane-air- additive mixtures were prepared by the method of partial pressures; ideal gas behavior was assumed. The pressures of propane, air, and addi- t i v e were read on a precision absolute manometer with the aid of a cathetometer.

The additives constituted only 0.5 percent of the total mixture by volume. This cqrresponds t o a partial pressure of 3.8 millimeters of mercury i n a mixture prepared a t a total pressure of 1 atmosphere. Therefore, i n order t o avoid the possibi l i ty of large error in the aadi- tive concentration, a 5.0 percent mixture of additive and air was f i r s t prepared. This m i x t u r e w a s then admitted to the s torage carboy t o a partial pressure of 60 millimeters of mercury; propane was added t o the desired partial pressure; and f inal ly dr ied air was l e t in to b r ing the total pressure to 600 millimeters. The concentratfon of additive in the resultant m i x t u r e w a s thus 0.5 percent by volume, based on the t o t a l mixture. This procedure could not be followed when chloropicrin (CC13NO~) was added, because it has a low vapor pressure; it w a s therefore neces- sary to measure the partial pressure corresponding t o 0.5 percent of t h i s additive directly on the manometer. Some precision was thereby sacri- f iced, but it i s believed that the chloropicrin concentration w a s within 11.3 percent of 0.5 percent by volume.

The a i r used t o make up the mixtures was passed through Ascarite ( t o remove carbon dioxide) and Anhydrone ( t o remove water vapor). The dew point of the dried air was found t o be less than -36' Fj it there- fore contained no m r e than 0.03 percent water vapor.

The pressure limits of binary mixtures of propane and air and of sdaitive and sir were also determined. These mixtures were prepared i n the storage carboy according to t he procedure described i n reference 5.

Experimental procedure. - The tests were carried out in the manner described i n reference 5. Most of the pressure limits were established to within fl millimeter of mercury. That is , two pressures were found that differed by 2 millimeters, the higher of which permitted flame prop- agation throughout the 3.73-centimeter tube, whereas the lower caused extinction a t the mouth of the tube. The limit recorded was the average of the two pressures.

Very few cases of e r r a t i c flame behavior were noted. The ones that were observed occurred with very lean mixtures, i n which the flame was sometimes extinguished between the mouth of the 3.73-centimeter-diameter t&e and i t s upper end, and with some r ich mixtures of carbon disulfide and air. Even in these cases, however, the reproducibility of the pres- sure limits was good.

c

NACA RM E53129

Analysis of Data

5

A curve of pressure limit for flame propagation in a 3.73- centimeter-diameter t&e against volume percent propane in t he propane- additive-air mixture w a s plotted for each adaitive. A reference curve for binary propane-air mixtures was d s o determined (fig. 2). ~n some cases, the mode of action of the additive could be determined by simple comparison of these two curves; the presence of the additive caused a displacement of the curve for ternary mixtures, w i t h relation t o the reference propane-air curve. The comparison was faci l i ta ted by plot t ing both curves on a single graph; examples may be seen i n f igme 3.

In the case of contbustible additfves, the propane-additive-air mix- tures contained 0.5 volume percent more fuel than is indicated by the . volume percent of propane. In such mixtures, then, the lean limits would be expected to be broadened, that i s , t o l i e a t lower prapane concentra- tions than for propane alone. Conversely, the rich limits should be narrowed, since the added conibustible contributes i t s rn oqygen demand in addi t ion to tha t of propane.

This method of presenting the data permits Immediate detection of flve types of additive action by simple coqarison of the limit curve fo r mktures of propane, air, and additive with the curve for propane and air:

(1) If the rich side of the additive curve lies outside (i.e., t o the right of) the reference propane-air curve, the adative exerts a definite promoting influence on flame propagation in rich mixtures. This is t rue whether the additive i tself is conktustible o r not, because even an inert additive should narrow the r ich limit by replaclng some of the oxygen.

(2) If the additive is combustible and the lean side of the additive curve lies inside (i.e., t o the right of ) the reference propane-air curve, the additive exerts a definite inhibiting influence on flame propagation in lean mixtures.

(3) If the additive is incombustible and the lean side of the d i - tive curve lies outside (i.e., to the left o f ) the reference propane-air cwve, the additive exerts a definite promoting effect on flame propa- gation in lean mixtures.

(4) If the presence of the addi t ive increases the minimum pressure for flame propagation as compared with the rninFrmrm of the reference propane-air curve, the additive i s an inhibitor of flame propagation i n mixtures corresponding t o the mindmum.

6 NACA RM E53129

(5) Similarly, if the presence of the additive decreases the mini- mum pressure f o r flame propagation, the additive acts as a promoter.

In four cases, decision as t o t h e mode of action of the additive cannot be made without further analysis of the data; these cases are:

( 6 ) If the additive i s conibustible and the lean side of the additive curve l ies outside the reference propane-air curve, the additive may be acting either as a promoter or as an added fuel. M tn

0 M (7) If the additive i s incombustible and the lean side of the addi-

t i ve curve l ies inside the reference curve, the additive may be acting ef ther as an inert gas or as a chemical inhibitor of flame-propagation reactions.

(8) If the additive i s conibustible and the r ich portion of the addi- t i ve curve l ies inside the reference curve, the additive may be acting as added fuel or as a chemical inhibitor.

(9) If the additive i s inconibustible and the r ich portion of the additive curve l ies inside the reference curve, the additive may be act- ing a s an iner t gas or as a chemical inhibitor.

The instances i n which the additive is combustible and causes both lean and r ich limfts t o occur a t lower prapane concentrations (cases (6) and (8)) were analyzed by means of the l a w of Le Chatelier, or nrLxture rule. The object was t o determine whether the displacement of the limits could be accounted fo r wholly by the contribution of the additive t o the fuel content of the mixture.

The mixture rule was formulated t o deal with the flammability limits of mixed fuels at atmospheric pressure (ref. 7 ) . It s ta tes tha t a simple additive relation exists between the proportions of t he fue l s i n a lean- limit mixture, as expressed in the followfng equation:

where

%, n2

N1

N2

percentages of each gas i n a lean-limit mixture of the two i n air, at atmospheric pressure

percentage of f i r s t gas at lean limit i n air, a t atmospheric pressure

percentage of second gas at lean llmit in air , a t atmospheric pressure

NACA RM E53129 7

w

w 8

The assumptions that the fuels burn similarly and that they do not in- terfere with one another are inherent in equation (1). The W i d i t y of equation (1) ha8 been tes ted for a large nuniber of mixtures a t atmos- pheric pressure (ref. 7 ) . It has been found that the r ich llmfts as well as the lean limits obey t h i s rule i n many instances. However, large deviations are solnetimes found, and these &e indicative of depar- ture from the assumptions of similar and independent conibustion of the two fuels.

In the present work, the mixture r u l e was applied t o the low-pressure limits of flame propagation i n propane-air mixtures containing 0.5 per- cent conibustible additive rather than to the flammability limits a t at- mospheric pressure. In equation (1) , n2 was set equal t o 0.5, the addi- tive concentration; and the va lues of N1 and Nz were read from curve6 of pressure limit against volume-percent propane or additive in air a t a given pressure. The expression was solved fo r nl, the concentration of propane i n a mixture of propane, additive, and air whose pressure limit of flame propagation in a 3.73-centimeter-diameter tube is the specified pressure. The calculated value of nl was then compared with the ex- perimentally observed value.

Fropane-Air Mixtures

It w a s first necessaryto ascertain whether the pressure limits of prqeane-air mFxtures determined i n the modlfied flame t&e agreed with previous results. The curve of preseure Ilmft against volume percent of propane i n air i s presented in figure 2. It is interesting t o note that the curve does not contain irregulas lobes on the r i c h side such as were found in the course of work on the effects of flame-tube diameter on pressure limit (ref. 5). The flame tubes used in the previous investi- gation differed from the one used in the present investigation i n two ways : (1) the propagation section m s u30 centimeters long, instead of 50 centbeters ; (2) the juncture of the ignition and propagation sections was tapered, instead of abrupt. The same capacitance-spark ignition source was used i n both cases.

It should be emphasized that the irregular lobes described in ref- erence 5 do not correspond to the lobes reported in reference 8 and ascribed to t he occurrence of cool flames. The cool-flare lobe reported in reference 8 f o r propane appeared i n mixtures richer than 7.2 percent propane by volume (the richest mixture studied in the present work con- tained 7.28 percent propane). The i r regular i t ies shown in reference 5 appeared i n mixtures leaner than 7.2 percent propane. It was therefore concluded thak the anomalies in reference 5 were probably due to the effects of aerodynamic disturbances on the flame front.

8 NACA RM E53129

It i s not known which of the two modifications in flame-t&e design was responsible f o r the elimination of the irregularities reported in reference 5 and the resulting smooth curve shown in f igure 2. 'In any event, the present curve is similar to CUTVeG of fuel concentration against quenching distance measured by the minimum s l i t width for f lash- back of a Bunsen flame (ref. 9) . The correspondence of t h e c r i t i c a l tube diameters for flame propagation, obtained from pressure-limit mea- surements, and the minimum slit widths of reference 9 is established in reference 5. It is shown that the s l i t widths are about 0.7 times the c r i t i c a l tube diameters for lean propane-air mixtures. With the present 8 data, this relat ion is also found to. hold for somewhat r ich mixtures. M The following table shows values of the r a t io of minimum sl i t width fo r flash-back t o c r i t i c a l %me diameter (3.73 cm) at corresponding pressures -

and propane concentrations.

Pressure, m i n i m width f o r percent mm Hg

Ratio of Minimum s l i t Propane,

by sl i t width f lash-back, volume Gi t o c r it i c a l

( a) diameter (a)

70

.71 2.63 5.50 54

.66 2.47 5.00 47

.66 2.47 '4.03 43

.62 2.33 3.50 52 0.66 2.45 3.00

aFig. 2. bRef. 9. CStoichiometric .

The values i n this table me within the range of values, 0.49 to 0.78, given i n reference 10 for the ratios of slit width t o tube diameter fo r flash-back of a Bunsen flame i n propane-air mixtures over the same range of concentrations. This resul t once again emphasizes that the low- pressure limits of flame propagation of propane-air mixtures may be gov- erned by quenching.

Propane - A i r -Additive Mixture s

Curves of pressure limit i n a 3.73-cent-ter-diameter t&e against volume percent propane, f o r the mixtures containing 0.5 volume percent of additive we presented in figure 3. The propane concentration is based on the t o t a l mixture. Each curve is accompanied by the curve f o r propane and air alone, so tha t the e f fec t of the additive on the low- pressure limits of flame propagation may be readily 6een. -

NACA RM E53129 9

Propane, a i r , and 0.5 percent ethylene. - Figure 3(a) shows the effect of 0.5 percent ethylene on the low-pressure l imits of flame propa- gation of propane-air mixtures. Ethylene was chosen because it was ex- pected t o a c t simply as an added hydrocarbon fuel, and the curve should serve as a basis for comparison t o detect promoting o r inhibiting effects of other coIribustible additives. Figure 3(a) shows that, as anticipated, the lean limits are broadened and the r i c h limits are narrowed by the presence of the added conibustible.

5!! 8 w Propane, air, and 0.5 percent ethyl nitrate ( C a 5 O N O 2 ) . - It has

been reported that e thyl n i t ra te broadens the r ich limit of butane i n air at atmospheric pressure (ref. 3). Inasmuch as e thyl n i t ra te i s a conibustible, it should have the opposite effect, as does ethylene. The o m e n contained i n the molecule fs not sufficient to oxidize the added ethyl ni t ra te to carbon monoxide and water, let alone sufficient to pro-

concluded that ethyl ni t ra te acts as a flame promoter i n r i c h butane- air mixtures. Figure 3(b) shows that this material also promotes propa- gation in rich propane-alr mfxtures a t reduced pressures; the pressure- limit curye f o r propane, air, and ethyl ni t ra te lies outside the propane-air curve on the- r ich s ide .

cu vide extra oxygen f o r the conibustion of butane. It must therefore be

&

Propane, afr, and 0.5 percent chloropicrin (~~13~02). - Ashmore and Norrish found that chloropicrin was a sensit izer f o r thermal explosions of hydrogen-oxygen and hydrogen-chlorine mixtures. Under some cfrcum- stances, however, it c d d also act as an inhibitor, presurmably because of the formation of nitrogen oxychloride (EToc1) by decomposition of the chloropicrin at higher temperatures (refs. l l and 12) . Figure 3(c) shows that chloropicrin promotes flame propagation i n r i c h propane-air mixtures; t h i s effect appears t o be quite strong. The I-lmits of lean mixtures, on the other hand, f a l l inside the propane-air curve. In these cases, chloropicrin seems t o act as an inert gas or as an inhibitor.

Propane, air, and 0.5 percent hydrogen. - The flammsbility charac- t e r i s t i c s of hydrogen as a fuel and as an additive fn mixtures at atmos- pheric pressure are described in reference 7. The composition 1Mts of hydrogen i n air at atmospheric pressure are unusually broad, 4.0 t o 75 percent by volume. Such an eas i ly flammable fuel might be expected t o exert a promoting effect on flame propagation when added t o hydrocarbon- air m i x t u r e s . Nevertheless, it was found that the effect of added hydro- gen on the lean limits of the saturated hydrocarbons methane and ethane at atmospheric pressure is simply that of an added fuel. In the case of the unsaturated compound, ethylene, hydrogen inhibits flame propagation in lean e r n e s .

The effects of hydrogen on the llmits of propane-air mixtures a t reduced pressures are sham by figure 3(d). It is seen that the addition of 0.5 percent hydrogen broadens the limits of lean mixtures slightly, as

10 NACA RM E53129 - w o u l d be expected i f the hydrogen acted as added fuel. In r f ch mixtures, the additive curve, as drawn , a t first l i e s just outside the propane-air curve, f r o m approximately 4.5 t o 6.5 percent propane. I n view of the experimental uncertainty in the pressure-limit measurements, it would perhaps be more correct t o s t a t e that the additive curve v i r tua l ly coin- cides w i t h the reference curve i n this concentration range. In e i ther case , hydrogen behaves as if it were promoting flame propagation in these mfir+ures; it is an added fuel and should therefore cause the r ich side of the additive cwve t o f a l l inside the propane-air reference curve. The additive curve crosses the propane-air curve a t about 6.5 percent and l ies within it fo r richer mixtures; that is, the hydrogen behaves as if it were an added fuel in these mixtures.

! m

Propane , a i r , and 0.5 percent hydrogen s u i d e . - Figure 3(e) shows that 0.5 percent hydrogen sulfide affects the pressure limits of propane- air mixtures qualitatively in much the 8ame way as aoes hydrogen.

Propane , a i r , and 0.5 percent carbon disulfide. - The appearrance of the additive curve in f igure 3(f) is qualitatively almost the same as i n

Propane, air, and 0.5 percent methyl bromide. - Recent experiments have shown that 0.5 percent methyl bromide narrows the composition range of flammability at atmospheric pressure i n the case of ethylene, methane, and n-hexane (ref . 13). Figure 3(g) shows that the same effect holds with-prapane-air mixtures a t reduced pressures. The entire additive curve lies inside the propane-air curve; this indicates that methylbro- mide inhibits flame propagation i n both rich and lean mixtures. I n addition, the minimum pressure for flame propagation i s markedly in- creased, from 42 t o 51 millimeters of mercury. Methyl bromide w a s the only additive tested that had a definite influence on the m i n i m of the pressure- lWt curve. Under certain conditions, mixtures of methylbro- mide and air are capable of propagating flame a t atmospheric pressure within m o w concentration limits If a very strong source i s provided to igni te the flamtmXble mixtures (ref. 7). Thus, although th is additive is capable of acting as an added fuel , the lean side of the curve in f i g - ure 3(g) shows that it does not do so i n propane-air mixtures.

-

Pressme l b i t s of additive-air mlxbures. - The p r e v i a section described qualitatively the effects of seven additives on the low-pressure limits of flame propagation of propane-air mixtures. Definite statements as t o promoting o r inhibiting action were confined t o cases in which the r ich o r lean sides of the additive curves lay t o the right of the corres- ponding lids of the propane-air reference curve. The Instances in which the additive i s colnbustible and causes both lean and r ich limits t o occur -

NACA RM E53129 11

w

w %

c': 5

a t lower propane concentrations remain t o be discussed. It is of inter- es t t o determine whether the displacement of the limits can be explained wholly by the contribution of the additive t o the fuel content of the mixture. This w a s done with the aid of the mixture rule, equation (1)

The d u e s of Nl were obtained from f igme 2. The pressure limits of the aaditives in a,ir had not previously been determined by the present technique; it was therefore necessaryto measure them i n order to obtain the desired values of N2. These determhations were made f o r mixtures of ethylene, ethyl nitrate, hydrogen sulfide, and carbon disulf ide in air; the curves of pressure lhit egainst volume percent additive in air are presented i n figure 4. With the exception of the ethylene-air curve, the r ich sides of the curves are not complete because of phy-sical limi- tations. In the case of e thyl ni t ra te , the vapor pressure at room tem- perature limited the concentration that could be obtained; in some r ich mixtures of hydrogen sulfide w i t h air and carbon disulfide w i t h air, ignit ion difficult ies were encountered w i t h the capacitance spark. The two-lobed curve f o r hydrogen sulfide (fig. 4(c)) was the only one of this type observed in t h i s investigation.

Chloropicrin and methyl bromide were considered nonflammable, and no attempt was made t o measure pressure limits.

One pressure lfmit was determined f o r a mixture of 6.89 percent, by volume hydrogen i n air. The limit found was 77 millimeters of mercury.

had t o be observed i n a completely darkened room after the eyes had be- come somemat adapted t o the darkness. It was a l so necessary to shield the eyes from the b r i l l i an t flash of the ignition spark. The.lean l i m i t of hydrogen i n air a t 1 atmosphere i s 4.0 percent (ref. 7) . The pres- sure lFmit of 30 percent hydrogen fn air i n a 3.73-centimeter-diameter tube was estimated from tb quenching-distance data of reference 1 by use of the relation between quenching distance and cr i t i ca l tube diameter for f l q e propagation pointea out in reference 5. This limit was esti- mated t o be 7 t o 8 millimeters of mercury. In view of the eqe r imen td d i f f i cu l t i e s i n the measurement of hydrogen-air pressure limits, the matter w a s not pursued further, and these three points were taken t o de- fine the pressure-limit curve i n an approximate fashion. The data are sumarized in the following table:

- Flames propagating i n this lean m i x t u r e were virtually nonluminous; they

Hydrogen i n air, 3.73-cm-diam. tube, percent by volume Pressure limit i n

mH@; 4.0 760 (ref. 7) 6.89 77 (measured)

30 7-8 (estimated from data of ref . 1)

12 NACA RM E53129

Comparison of calcu3ated and observed pressure limits of propane, air, and 0.5 percent additive. - The observed pressure limits of propane- air-additive mixtures are compared i n figure 5 w$th those calculated by the mixture rule. The calculated limits are shown by the sol id curves. The observed limits me Shawn by the curves of figure 3 which are in- serted as dashed l ines for comparison. The calculated limit curves were constructed from values of nl, the percent propane i n the lean-limit mixture of propane, air , and 0.5 percent additive, calculated by means of equation (1). Comparison of the lean sides of the calculated curves w i t h the observed limits in f igures 5( a) t o (d) shows tha t the mixture rule holds quite well at reduced pressures for lean-limit mixtures of propane and air containing 0.5 percent by volume ethylene, ethyl nitrate, hydrogen, or hydrogen sulfide. The best agreement between experiment and the predictions of the mixture rule i s shown by the mixtures con- ta ining e thyl ni t ra te o r hydrogen. The curves for the observed pres- sure limits of lean mixtures containing ethylene or hydrogen su l f ide l i e uniformly just inside the calculated curves. The deviations i n limit concentration are small along the more vertical portions of the curves and become fairly definite near the minimum (figs. 5(a) and (a)). Some slight inhibitory action of ethylene and hydrogen sulfide on flame prop- agation in lean propane-air mixtures may be indicated. However, none of the four additives had any s t r iking effect , either of promotion or in- hibition3 and the broadening of the lean limits noted in f igures 3(a), (b), (a), an& (e) is seen t o be very nearly explainable by the contribu- t i on of the addi t ives to the fuel content of the mixtures.

The observed and predicted Umits of lean mixtures of propane, air, and 0.5 percent carbon disulfide axe compared in f igure 5(e). The ex- perimental curve l ies far wLthin the predicted curve; the addition of 0.5 percent by volume carbon disulfide thus appears t o have a strong inhibitory effect on the pressure limits of lean propane-air mixbures. The l imits of mixtures of carbon disulfide with ether, benzene, acetone, and acetaldehyde at atmospheric pressure do not obey the mixture rule (ref . 7 ) .

The magnitude of the inhibitory effect af carbon disulfide may be determined with the aid of the mixture rule. Values of nl, the percent propane in lean-limit mixtures of propane, air, and 0.5 percent carbon disulfide, were read from the pressure-limit curve, figure 3( f ) . These

values were used t o compute the B L ~ I - N1 " Eo n1 n2 According to equation (l),

t h i s sum equals unity i f the mixture behaves i d e w . If the sum i s greater than unity, inhibition i s indicahed in the case of lean-limit mixtures; tha t is, the l i m i t mixture must contain more propane than predicted by the mixture rule i n order for the flame t o propmate. In order t o emphasize that carbon disulfide has an inhibitory effect in lean propane-air mixtures, the values of the sum were subtracted from unity t o give negative numbers.

UCA RM E53129 13

On the assumption that the slight but W o r m discrepancies between calculated and observed lean limits s h m by ethylene and hydrogen sul- fide (figs. 5(a) and (a)) indicate inhibition, the preceding calculation was also carried out for these two additives. The resul ts axe summmized in table I. Ethylene and hydrogen sulfide are seen t o inhibit flame propagation at reduced pressures in lean mixtures t o the extent of 3 t o 4 percent. The inhibition due t o carbon disulfide i s much larger, of the order of 20 percent.

Reference 14 reported that 0.5 percent hydrogen sulfide in propane- air m€xtures inhibited the maximum burning velocity by about 3 percent; the effect was evaluated by means of a mixture rule analogous t o the one used i n this investigation. In the case of th is additive, therefore, the effects on both conibustion properties are in accord. However, ref- erence 14 showed tha t 0.5 percelrt ethylene i n propane-air mixtures did obey the mixture rule; the present results do not agree with this con- clusion. On the other hand, recent studies of the s tab i l i ty of flames i n mixtures of propane, ethylene, and air indicated that the mixture rule i s not always obeyed by these fuels (ref. 15). A sllght suggestion of inhibition was reported. The possible significance of the calculated fnhibitory effects of ethylene and hydrogen sulfide, as presented i n table I, therefore remains in doubt.

Ethylene and hydrogen sulfide were the only t w o additives for which the necessary data were obtained to tes t the applfcat ion of the mixture rule t o r i c h mixtures. Figures 5(a) and (a) show that the predicted rich limits correspond t o the experimental ones reasonably well in view of the fact that the mixture rule was originally intended t o apply t o lean limits.

No attempt was mae t o calculate the pressure-limit curves in the region close to the minimum, because of two diff icul t ies : ( I ) The mini- m u m of the propane-air 'ana additive-afr limit curves a;td not occur a t the sane pressure; (2) the minimums did not occur at the same percentage of stoichiometric. Therefore, uncertainty existed as t o the proper values of N1 and 'N2 t o choose for the calculation.

Evaluation of effects of chloropicrin and methyl bromide on pressure limits of propane-air mixtures. - Figures 3(c) and (g) show that chloro- picr in and methyl bromide, respectively, have a definite inhibitory effect on the pressure limits of lean propane-air mixtures, since the lean sides of the curves for these additives l i e t o the right of the reference propane-air curve. It seems reasonable t o assume that both of these addi- t ives may be considered inconibustible; thus, their action cannot be due t o a greater aff ini ty for oxygen than propane possesses. Tu0 possibil i- t i e s remain:

14 NACA RM E53129

(1) The additive may act simply as an inert gas that replaces part of the oxygen i n the mixture.

( 2 ) The additive may ac t in a specific chemical manner t o inhibit the combustion reactions of propane.

The first of these possibi l i t ies w a s checked by measuring the pres- sure limits i n propane-air mix tu res containing 0.5 percent inert gas. Pure nitrogen (99.9 percent) was chosen. The results are presented in figure 6, together with the reference propane-air curve. It w i l l be seen that the added nitrogen has no appreciable effect on the pressure limits of any mixture studied. The curves for propane, air, and 0.5 percent chloropicrin and for propane, a i r , and 0.5 percent methylbro- mide are also reproduced in f igure 6 . Comparison shows that the effects of these additives are much greater than tha t of nitrogen. It might be argued that the heat capacities of chloropicrin and methyl bromide are considerably greater than the heat capacity Of nitrogen, so they should be more effective inerts. However, in view of the low additive concen- t ra t ion, it is believed that this is an unimportant constderatLon. It is therefore concluded that chloropicrln acts in a specific chemical manner t o inhibit flame propagation a t reduced pressures i n lean propane- air mixtures (although it is a promoter in rich mixtures); methyl bro- mide has a chemical inhibitory effect i n both lean and r ich mFxtures .

i

SUMMARY OF RESULTS

Seven additives in 0.5-volume-percent concentration were studied t o determine their effects on the low-pressure limits of flame propagation of propane-air mixtures, with the following results :

1. of the seven additives tested, none was found t o promote flame propagation in lean mixtures at reduced pressure more than could be ex- plained by the contribution of the additive to the total fue l i n t he mixture.

2. No additive was found that significantly lowered the minimum pressure for flame propagation.

3. Two additives - ethyl ni t ra te and chloropicrin - were definite promoters of flame propagation i n r ich propane-air mixtures.

4. Of the Combustible additives tested, ethylene and hydrogen su l - f ide appeared to inh ib i t flame propagation in lean mixtures t o a s l ight degree. Carbon disulfide was a marked inhibitor.

5. Both chloropicrin and methyl bromide inhibited flame propagation in lean mixtures. Methyl bromide also increased the minimum pressure f o r flame propagation and inhibited propagation i n r i c h mixtures. The effects were too large to explain in terms of dilution by an iner t gas.

NACA RM E53129

C ONCUTSIONS

15

The following conclusions may be drawn from t h i s study:

1. The Le Chatelier law may be applied to t he limits of flame prop- agation at reauced pressures for some mixed fuels.

2. Chloropicrin has a chemical inhibitory action on flame propaga- t i on i n lean propane-air mixtures a t reduced pressures, even.though it is a promoter i n r i ch mixtures. Methyl bromide i s a chemical inhibitor i n both lean Rnd r ich mixtures.

3. Carbon disulfide inhibits flame propagation i n lean propane-air mixtures t o the extent of about 20 percent, as shown by deviations from Le Chatelier ' s l a w .

Lewis Flight Propulsion Laboratory National Advisory Committee f o r Aeronautics

Cleveland, Ohio, October 1, 1953

1. Lewis, Bernard, and von Elbe, Guenther : Combustion, Flames, and Explosions of Gases. Academic Press, Inc., 1951.

2. Leason, D. B.: The Effect of Gaseous Additions on the Burning Veloc- i t y of Fropane-Air Mixtures. Fourth Symposium (International) on Combustion, The WflLiaw&Wilkins Co. (Baltimore), 1953, pp. 369-375.

3. Egerton, Alfred, and Powling, J. : The L imi t s of Flame Propagation at Atmos heric Pressures. I. The Influence of "PTomoters." Proc. Roy. SOC. ?London), ser. A, vol. 193, May 27, 1948, pp. 172-190.

4. Dixon-Lewis, G., and Linnett, J. W.: The Effects of Organic Sub- stances on the Upper Limits of Inflammability of Some Eydrogen- Carbon Monoxide-Air Mixtures. Proc. Roy. SOC. (London), ser. A, vol. 210, Dec. 7, 1951, pp. 49-69.

5. Belles, Frank E., and Simon, Dorothy M. : Variation of the Pressure Limits of Flame Propagation with Tube Diameter for Propane-Air Mix- tures. NACA RM E51J09, 1951.

6. Simon, Dorothy M., and Belles, Frank E.: An Active Particle Diffusion Theory of Fl8n-e Quenching f o r hinar Flames. W A RM E5lLL8, 1952.

7. Coward, H. F., and Jones, G. W.: Limits of Flammability of Gases and Vapors. B u l l . 503, Bur. Mines, 1952.

16 NACA RM E53I29

8. DiPiazza, James T., Gerstein, Melvin, and Weast, Robert C. : Flamma- b i l i t y L i m i t s of Hydrocarbon-Air Mixtures. Reduced Pressures. Ind. and Eng. Chem., vol. 43, no. 12, Dec. 1951, pp. 2721-2725.

9. Friedman, Raymond, and Johnston-, W. C. : The Tall-Quenching of Laminar Flames as a Function of Pressure, Temperature, and A i r - F u e l Ratfo. JOW. Appl. Phm., V O ~ . 21, no. 8, A x . 1950, pp. 791-795.

10. Harris, Margaret E., G r u m e r , Joseph, von Elbe, Guenther, and Leuis, VI Bernard: Burning Velocities, Quenching, and Stab i l i ty Data on Nonturbulent Flames of &thane and Propane with Oxygen and Nitro- gen. Third Symposium on Combustion and Flame and Explosion Phenomena, The Williams bWilkins Co. (Baltimore), 1949, pp. 80-89

(D 0 M

11. Ashmore, P. G., and Norrish, R. G. W.: A Study of Sensitized Explo- sions. V I I I . Experimental Work on the Hydrogen-Oxygen Reaction Sensitized by Chloropicrin. Roc. Roy. SOC. (London) , ser. A, V O ~ . 203, Octo 24, 1950, pp. 454-471.

12. Ashore, P. G., and Norrish, R. G. W.: A Study of Sensitized Explo- sions. M. Experimental Work on the Sensitization of Ignitions of Hydrogen-Chlorine Mixtures by Chloropicrin. Roc. Roy. SOC. (London), ser. A, vol . 203, Oct. 24, 1950, pp. 472-486.

13. Burgoyne, J. H., and Williams-Lee, G. : The Influence of Incombstible Vapours on the Limits of Inflammability of Gases and Vapours i n A i r . Proc. Roy. SOC. (London), ser. A, vol. 193, July 21, 1948, pp. 525-539.

14. Kurz, Philip F. : The Role of Additives i n the Conibustion Reactfons of Hydrocarbons. I. The Influence of Hydrogen Sulfide on the Flame Speed of Prapane-Air Milrtures. Tech. Rep. No. 15036-5, Battelle Memorial Inst., June 16, 1952. (Contract No. AI? 33( 038) - 12656, E.O. NO. 460-35 S.R. -8.)

15. KWZ, Phi l ip F. : The Stabi l i ty of Laminar Propane-Ethyl-ene Flames Abs. of papers presented at Am. Chem. SOc meeting, Chicago (111. , Sept. 5-11, 1953.

NACA RM E53129

f Pressure, I Aadltive 1

150 I 140 130 120 110 100 90 80 70 60 50

I Ethylene, I Hydrogen sulfide, I Carbon disulfide, I

-0 -04 - -04 - -03 - -03 - .02 - .02 - .02 - .02 - -03 - -04 - -07

-0.03 - -03 - -03 - -04 - -04 - .04 - -03 - .03 - -06

-0.22 - .22 - -21 - .21 - .a - .20

17

NACA RM E53129

To mirture storage and Vaouum

IL 8.7

Figure 1. - Flame tube for measurement of low- preeaure limita of flame propagation.

W A RM E53129

50 75 100 125 150 I35 200 Propane in aLr, percent stoichiometric

Propane in air, percent by volume

Figure 2. - Pressure limits of flame propagation of propane- air mixtures in 3.73-centimeter-diameter tube.

20 NACA RM E53129

IUCA RM E53129

40 6oE 2

\ I

\ \ \ \

3

21

,-c-. Propane, air, and ethyl nftrete . "- - -Propane and a i r

4 5 6 7 8 9 Propane in mixture, percent by volume

(b) Hhyl nitrate, 0.5 volume percent.

Figure 3. - Continued. Effects of additives on low-pressure limits of flame propagation in propane-air mixtures in 3.73-centimeter- diameter tube.

22

200

160

140

120

100

80

60

40 2 3

NACA RM E53129

(c) Chloropicrin, 0.5 volume percent.

Figure 3. - Continued. Effects of additive8 on luw-presmre llmite of flame propagation in propane-air mixtures in 3.73-centimeter- diameter tube.

w 0 cn w

4 5 6 7 Propane in mixture, percent by volume

8 9

NACA RM E53129

2 3 4 5 6 7 Propane in mixture, percent by volume

(a) Hjrdrogen, 0.5 volume percent.

a

23

Figure 3. - Continued. Effects of additives on lm-pressure m t s of flame propagation in propane-air mixtures in 3.73-centimeter- diameter tube.

24

200

160

60

40

NACA RM E53129

Propane in ruixtw, percent by volume

(e) mdrogen sulfide, 0.5 volume percent.

Figure 3. - Continued. Effects of additives on low-pressure limits of flame propagation in propane-air mixtures in 3.73-centimeter- diameter tube.

.

NACA RM E53129 25

220

200

60

40

Propane in mixture, percent by volume

( f ) Carbon disulfide, 0.5 volume percent.

Figure 3. - Continued. Effects of additives on low-pressure limits of flame propagation in propane-alr mixtures in 3.73-centimeter- diameter tube.

26 NACA RM E53129

(g) Methyl bromide, 0.5 volume percent.

Figure 3. - Concluded. Effects of additives on low-pressure lMts of flame propagation i n propane-& mixtures Ln 3.73-centimeter- diameter tub e.

.

3 0 to

KACA RM E53129 27

mhylene in air, percent by volume

50 75 100 125 150 175 200 225 250 275 300 Ethylene i n air, peroent atoichiometric

Figure 4. - Low-pressure Units of flame propagation in additive-air mixtures in 3 .73-cent imeter -der tae.

28 NACA RM E53129

Ethyl nitrate in air, percent by volume

50 75 100 125 150 E t h y l nitrate in air, percent stoichiometric

(b) E t h y l nitrate - air mixtures.

Figure 4. - Continued. Lar-presaure limits of flame propagation in additive-air mixtures in 3.73-centimeter-diameter tube.

NACA RM E53129 29

EIydrogen sulfide in a h , percent by volume L I I I I I I 50 75 100 125 E O 175 200

Hydrogen sulfide i n aFr, percent etoichiometrlc

(c) ~tydrogen suide - air mixhmas.

Figure 4. - Continued. Low-preeeure limits of flame propagation in additive-air rd.xLures fn 3.73-centimeter-dla&er tube.

30

100 -

80 -

60 -

-

a-

-

20 -

-

- 0

HACA RM E53I29

4 a 12 16 20 Carbon disulfide in air, percent by volume

24

~~ ~

50 100 150 200 250 300 350 400 450

Carbon disulfide in air, percent stofchiometric

(a) Carbon disulfide - air mixtures. Figure 4. - Concluded. Low-pressure limits of flame propaga-

t i on in additive-air mixtures in 3.73-centimeter-diameter tube.

2 3 4 5 6 7 Propane i n mixture, percent by volume

(a) Ethylene, 0.5 volume percent.

31

Figure 5. - COTnparison of calculated and observed pressure limits of flame propagation of m i x t u r e s of propane, air, and 0.5 percent additive.

32 NACA RM E53129

2 3 4 5 6 7 8 Propane fn mixture, percent by volume

(b) Ethyl nit rate , 0.5 volume percent.

Figure 5 . - Continued. Comparison of calculated and observed pressure limits of flame propagation of mixtures of propane, air, and 0.5 percent additive.

IIACA RM E53129 . 33

2 3 4 5 6 7 8 Propane in mixture, percent by volume

(c) Hydrogen, 0.5 volume percent.

Figure 5. - Continued. C o m p a r i s o n of calculated and observed preemre limits of flame propagation of mixtures of propane, air, and 0.5 percent additive.

34 NACA RM E53129 .

Propane in mixture, percent by volume

(d) Hydrogen sulfide, 0.5 volume percent.

Figure 5. - Continued. Comparison of calculated and observed pressure limits of flame propagation of mixtures of propane, air, and 0.5 percent additive.

NACA RM E53129 35

.

(e) Carbon disulfide, 0.5 volume percent.

Figure 5. - Concluded. Comparison of calculated and observed pressure limits of fleme propagation of mixtures of propane, a i r , md 0.5 percent additive.

36 NACA RM E53129

20c

1%

16C

&? B

i 12c

i! PI

1oc

8c

60

40

Propane In mixture, percent by volume

Figure 6. - Effect of 0.5 volume percent nitrogel1 on preasure limits of flame propagation of propane-air mixtures.

. '